Deadlock when using fix addbornforce with Allegro potential

[2297019091]

I encountered a deadlock issue when using LAMMPS (2 Aug 2023) with the Allegro potential. After applying the patch files from the Allegro repository and compiling successfully, the program hangs at the following point:

text

LAMMPS (2 Aug 2023)
  using 1 OpenMP thread(s) per MPI task
Reading data file ...
  orthogonal box = (0 0 0) to (11.898254 11.898254 12.912077)
  1 by 1 by 1 MPI processor grid
  reading atoms ...
  135 atoms
  read_data CPU = 0.003 seconds
Allegro is using input precision f and output precision d

The program freezes completely after this output. The issue occurs when using the fix addbornforce command in my input script. Without this fix, the calculation runs normally.

I applied the patch files from the Allegro repository by executing the following commands:

cd /pair_allegro
./patch_lammps.sh /lammps
cp ./extra/fix_addbornforce.cpp /lammps/src/
cp ./extra/fix_addbornforce.h /lammps/src/

Environment
LAMMPS Version: 2 Aug 2023
Operating System: Linux
MPI: Open MPI 4.1.2
PyTorch: 1.11 (CPU only)

Debug Attempts
Tested with single process (no MPI) - same issue
Without fix addbornforce, calculation works

Has anyone else encountered deadlock issues with fix addbornforce?

Are there known compatibility issues between Allegro and LAMMPS 2023?

What debugging approaches would you recommend for this type of hang?

Could this be related to PyTorch initialization or MPI communication?